1. Field
Methods and apparatuses consistent with exemplary embodiments relate to video encoding and decoding using image up-sampling.
2. Description of the Related Art
Conventional image encoding and decoding splits one picture into macroblocks to encode an image. Thereafter, inter prediction or intra prediction is used to prediction encode each of the macroblocks.
Inter prediction is a method of compressing an image by removing a temporal redundancy between pictures and has motion estimation encoding as a representative example. Motion estimation encoding predicts each block of a current picture by using at least one reference picture. A predetermined evaluation function is used to search for a reference block that is most similar to a current block within a predetermined search range.
The current block is predicted based on the reference block, and a residual block—generated by subtracting a prediction block generated as a result of prediction from the current block—is encoded. In this regard, to more accurately perform prediction, interpolation is performed on the search range of the reference picture, sub-pixels of a pixel unit—smaller than an integer pet unit—are generated, and inter prediction is performed based on the generated sub-pixels.